The Methylenetetrahydrofolate Reductase C677T Polymorphism and Risk for Late-Onset Alzheimer’s disease: Further Evidence in an Italian Multicenter Study

Abstract

Background: A functional polymorphism in the methylenetetrahydrofolate reductase (MTHFR) gene, namely C677T (rs1801133), results in increased Hcy levels and has been associated with risk of late-onset Alzheimer’s disease (LOAD). Many investigators reported association between rs1801133 and LOAD risk in Asian populations and in carriers of the apolipoprotein E (APOE) ɛ4 allele, but recent meta-analyses suggest a contribution also in other populations, including Caucasians and/or northern Africans.

Objective: To further address this issue, we performed a relatively large case-control study, including 581 LOAD patients and 468 matched controls of Italian origin. APOE data were available for a subgroup of almost 600 subjects.

Methods: Genotyping for rs1801133 was performed with PCR-RFLP techniques.

Results: In the total population, the MTHFR 677T allele (OR = 1.20; 95% CI = 1.01–1.43) and carriers of the MTHFR 677T allele (CT+TT versus CC: OR = 1.34; 95% CI = 1.03–1.73) resulted in increased LOAD risk. Similarly, in APOE ɛ4 carriers, we observed an increased frequency of MTHFR 677CT carriers (CT versus CC: OR = 2.82; 95% CI = 1.25–6.32). Very interestingly, also in non-APOE ɛ4 carriers, both MTHFR 677T allele (OR = 1.38; 95% CI = 1.03–1.85) and MTHFR 677TT genotype (OR = 2.08; 95% CI = 1.11–3.90) were associated with LOAD. All these associations survived after corrections for age, gender, and multiple testing.

Conclusions: The present results suggest that the MTHFR C677T polymorphism is likely a LOAD risk factor in our cohort, either in APOE ɛ4 or in non-APOE ɛ4 carriers.

Keywords

Alzheimer’s disease APOE folate homocysteine methylenetetrahydrofolate reductase MTHFR C677T

INTRODUCTION

Impaired one-carbon metabolism, which is reflected by increased circulating homocysteine (Hcy) levels coupled to reduced folates and related B-group vitamins, is frequently observed in late onset Alzheimer’s disease (LOAD) patients [1]. Indeed, high Hcy and low folates are regarded as LOAD risk factors [1 –5]. Folates are cofactors in one-carbon metabolism allowing the remethylation of Hcy to methionine, and methylenetetrahydrofolate reductase (MTHFR) is an important enzyme in this pathway as it catalyses the irreversible conversion of 5, 10-methylenetetrahydrofolate (5, 10-MTHF) to 5-methyltetrahydrofolate (5-MTHF), the methyl donor compound for Hcy remethylation [2]. A functional polymorphism in the MTHFR gene, namely C677T (rs1801133), results in alanine to valine substitution at position 222 in MTHFR protein [6, 7]. MTHFR works as a dimer which is stabilized by physiological levels of folates, and the mutant TT enzyme is less stable and prone to dissociate into monomers at 37°C, particularly under conditions of reduced folate bioavailability [8, 9]. As a consequence, the mutant TT enzyme has a resultant mean activity which is 40–50% lower than the wild type CC one [7], and increases Hcy levels in LOAD patients or healthy elderly individuals, particularly in those with a folate deficiency state [9 –11].

A total of more than 40 case-control studies have been performed in different ethnic groups searching for association of the MTHFR C677T polymorphism with LOAD risk [11 –15], and their meta-analysis revealed that the T allele is a genetic risk factor for the disease, particularly in Asian populations and in carriers of the apolipoprotein E (APOE) ɛ4 allele, which is the major genetic risk factor for LOAD [13]. The association of the MTHFR C677T polymorphism with LOAD risk in other populations than Asians, or in APOE ɛ4 non-carriers, is still controversial and only suggested by some of the most recent meta-analyses that collected an increasing number of available case-control studies [12 –14]. These inconclusive findings are largely due to the relatively low sample-size of the case-control studies available in the literature [12 –14]. Indeed, only two out of more than 40 case-control studies addressing this issue included more than 300 cases and controls each [11, 15], but most of them are limited to less or about 100 cases and controls [12 –14].

We performed the present study to better address this issue in Italian elderly subjects of Caucasianorigin. Overall, we collected more than 1,000 samples from three different neurology units in Italy, making the present case-control study, at best of our knowledge, the largest one performed thus far searching for association between the MTHFR C677T polymorphism and LOAD risk.

MATERIALS AND METHODS

Study population

The analysis was based on 581 LOAD patients and 468 healthy matched controls (Table 1), including the 378 LOAD and 305 control samples with available MTHFR C677T data previously described by us [11], and additional 203 LOAD and 163 control samples collected from 2013 to 2016 at the recruiting centers. LOAD patients and controls were recruited at the Department of Neuroscience of the Pisa University Hospital, at the Department of Neuroscience, Psychology, Drug Research and Child Health of the University of Florence, and at Oasi Maria SS Institute for Research on Mental Retardation and Brain Aging of Troina. Trained neurologists at the recruiting centers visited cases and controls before inclusion in the study. LOAD patients met the Diagnostic and Statistical Manual of Mental Disorders (DSMIV) and NINCDS-ADRDA criteria [16, 17]. According to disease age at onset (>65 years) and absence of a family history of dementia all the patients were assumed to be sporadic LOAD cases. Healthy volunteer subjects were matched to LOAD patients for age at sampling, gender, and ethnicity (all the recruited individuals were Caucasians of Italian descent from at least three generations) (Table 1). All the control subjects underwent a rigorous neurological examination before inclusion in the study, in order to exclude the presence of cognitive impairment (Mini-Mental State Examination score over 26) or any other kind of neurological disorder. Furthermore, control subjects were also investigated for their familial history of neurological disorders, and only individuals with no relatives who developed LOAD or related diseases were included in the study. The APOE genotype was known for 375 LOAD patients and 204 age- and gender-matched controls (Table 1), and this is due to the fact that APOE data were largely missing from the previously collected 378 LOAD and 305 control samples [11] for which DNA samples are no-more available. APOE genotyping was performed as previously detailed [18]. Each subject gave an informed and written consent for the inclusion in the study that received approval from the Ethics Committee of the Pisa University Hospital (Protocol number 3618/2012).

Genotyping

Genomic DNA was isolated from whole blood by means of the QIAamp^® Blood Mini Kit (Qiagen, Milan, Italy), and stored at –20°C until assayed. MTHFR genotyping was performed by means of PCR-RFLP technique as previously described [11] and digestion products were visualized after electrophoresis on a 3% agarose gel stained with ethidium bromide. Internal quality controls with confirmed genotypes were included on each digestion.

Statistical analyses

To verify that genotype frequencies were in Hardy-Weinberg equilibrium (HWE) we used the Chi-square (χ²) analysis. Odds ratios (ORs) have been calculated by means of logistic regression analysis and given with 95% confidence intervals (CIs). We calculated both crude ORs as well as age and gender adjusted ones. Bonferroni’s corrected p-values are provided. Statistical analyses were performed with the MedCalc v.12.5 software. We used the statistical package QUANTO 1.2.4.exe. and data [minor allele frequencies (MAF) and Odds Ratios (ORs)] from previous large studies and-meta analyses of MTHFR C677T in Caucasian LOAD [11, 14] to design a case-control study with enough a priori power (>80%) to detect ORs > 1.2 for MAF ranging from 0.42 to 0.46.

RESULTS

Table 2 shows MTHFR C677T allele and genotype frequencies in the total cohort of LOAD and control samples, as well as in the available subgroups of APOE ɛ4 (+) carriers and non-carriers (–). Genotype frequencies in controls conformed to Hardy-Weinberg expectations (χ² = 2.18). The MTHFR 677T allele was associated with increased LOAD risk in the total population (OR = 1.20; 95% CI = 1.01–1.43), and carriers of the MTHFR 677T allele (CT+TT versus CC) resulted at increased LOAD risk (OR = 1.34; 95% CI = 1.03–1.73). APOE ɛ4 (+) carriers were higher in LOAD patients than in controls (46.9% versus 19.1%; p = 0.0001) (Table 1). In APOE ɛ4 carriers we observed an increased frequency of MTHFR 677CT carriers (CT versus CC: OR = 2.82; 95% CI = 1.25–6.32). Very interestingly, also in the 199 LOAD and 165 controls who do not carry the APOE ɛ4 allele, both MTHFR 677T allele (OR = 1.38; 95% CI = 1.03–1.85) and MTHFR 677TT genotype (OR = 2.08; 95% CI = 1.11–3.90) were associated with increased risk for LOAD. All these associations survived after correcting for age, gender, and multiple testing in logistic regression analysis (Table 2). Furthermore, among LOAD patients the frequencies of MTHFR 677T alleles were similar between carriers (176 LOAD patients) and non-carriers (199 LOAD patients) of the APOE ɛ4 allele (46% versus 48.5%, p = 0.51).

DISCUSSION

In the present study, we observed that the MTHFR 677T allele represents a weak genetic risk factor for LOAD in our population, and this is in agreement with all the most recent meta-analyses of genetic association studies, overall reporting ORs ranging from 1.2 to 1.5 for this polymorphism, depending on the ethnic group or the genetic model under investigation [12 –14]. All those meta-analyses have confirmed that the MTHFR 677T allele is a LOAD risk factor in Asian populations [12 –14], but showed association or trends for association also in Europeans and/or in mixed populations [5 , 13]. Particularly, the most recent meta-analysis in the field reported an OR of about 1.3 for the MTHFR 677T allele with respect to the 677C one in Europeans and northern Africans [14], and previous meta-analyses showed association in mixed populations [12, 13] and a trend for association in Caucasians [13]. At best of our knowledge, the present is the largest case-control study addressing this issue, and together with recent meta-analysis papers [12 –14] confirms that the association between the MTHFR C677T polymorphism and LOAD risk is not limited to Asian populations.

Some investigators also questioned possible interactions between MTHFR C677T and APOE ɛ4 allele in LOAD risk, being the latter the main and most replicated LOAD risk factor [12 , 18]. There is consensus in the literature, from either relatively large case-control studies or meta-analyses [12 , 15], indicating that the MTHFR C677T polymorphism increases LOAD risk in APOE ɛ4 carriers, particularly for heterozygous 677CT or combined CT+TT carriers [13]. Therefore, the present data reporting association of the MTHFR C677T polymorphism and LOAD risk in APOE ɛ4 carriers, although limited by a few number of APOE ɛ4 carriers in controls, are in complete agreement with the overall literature in the field [12 , 15]. Less clear, however, is the association of the MTHFR C677T polymorphism and LOAD risk in non-APOE ɛ4 carriers [12, 13]. In this regard, a meta-analysis performed including only well designed case-control studies, selected with rigorous diagnostic criteria and limited to those without HWE deviations in controls, reported association of the MTHFR 677T allele with LOAD risk also in non-APOE ɛ4 carriers (OR = 1.21; 95% CI = 1.04–1.42) [12], while a broader meta-analysis that included all the available studies showed only a trend in non-APOE ɛ4-MTHFR 677TT carriers (OR = 1.27; 95% CI = 0.97–2.02) [13]. The major problem leading to uncertainty is that there are only a few case-control studies with combined MTHFR C677T and APOE data [12]. Particularly, only two of such studies are available in Caucasians, the first including 122 non-APOE ɛ4 LOAD carriers [19] and the latter only 24 [20], so that any additional data in non-APOE ɛ4 carriers is largely desired [13]. In this regard, the present study shows that, among the 375 LOAD subjects with available APOE data, there is no difference in MTHFR 677T allele frequencies between APOE ɛ4 carriers and non carriers (176 and 199 subjects, respectively, with MTHFR 677T allele frequencies of 46% in carriers and of 48.5% in non-APOE ɛ4 carriers, p = 0.51), suggesting no increased presence of the T allele in the first group. Furthermore, when the 199 non-APOE ɛ4 LOAD carriers were compared to 165 non-APOE ɛ4 controls, both MTHFR 677T allele (OR = 1.38; 95% CI = 1.03–1.85) and 677TT genotype (OR = 2.08; 95% CI = 1.11–3.90) resulted in increased LOAD risk. Overall, present data are indicative that a contribution to LOAD risk for this polymorphism cannot be factored out in those without the APOE ɛ4 allele. Very recently, it was observed that the MTHFR C677T polymorphism modulates the associations between blood-based and cerebrospinal fluid (CSF) biomarkers of neurodegeneration [21]. Particularly, in carriers of the MTHFR 677C allele, a significant association was observed between circulating levels of the apolipoprotein E protein (ApoE) and CSF levels of the amyloid-β (Aβ₄₂) peptide. However, this association was not observed in MTHFR 677TT individuals or after including the APOE genotype in the model [21]. These findings [21], together with present data, suggest the need of further studies to better understand the combined effects of MTHFR and APOE genotypes on circulating biomarkers of neurodegeneration, including circulating Hcy and ApoE levels and CSF levels of Aβ₄₂. In this context, additional studies, including prospective studies in the elderly, are also required to confirm the increased LOAD risk that we observed in MTHFR 677TT individuals that do not carry the APOE ɛ4 allele, and the contribution of this genotype to circulating markers of neurodegeneration in those individuals.

The interest in MTHFR as a candidate LOAD gene increased after the publication of a Nature Genetics paper containing the first systematic analysis of the AD-related genetic association studies and indicating MTHFR among the AD susceptibility genes [22]. A subsequent genome-wide association study (GWAS) in French individuals confirmed a nominal association between MTHFR variants and AD [23], and subsequent meta-analyses of the literature provided further evidence of an association between the MTHFR C677T polymorphism and LOAD risk [12 –14]. Very interestingly, another gene involved in tetrahydrofolate (THF) metabolism was associated to LOAD risk in a subsequent GWAS [24], and particularly the MTHFD1L gene, which encodes the methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 1-like protein catalyzing the reversible synthesis of 10-formyl-THF to formate and THF, an important step in Hcy conversion to methionine [24]. Additional studies confirmed an association between MTHFD1L polymorphisms and LOAD risk in Asian populations [25, 26], and both MTHFR and MTHFD1L polymorphisms are believed to contribute to LOAD risk by increasing circulating Hcy levels [5, 24]. Despite that MTHFR C677T is the major functional MTHFR polymorphism [7], additional variants are known in the MTHFR gene, and a haplotype containing the MTHFR 677C allele has been linked to reduced LOAD risk, particularly in non-carriers of the APOE ɛ4 variant [27]. However, additional studies are required to clarify the contribution of different MTHFR haplotypes to circulating Hcy levels in LOAD patients and their interactions with APOE variants.

Concerning the link between rs1801133 and LOAD risk, it is well known that it contributes to increased Hcy levels [5]. Indeed, many investigators observed increased Hcy levels in LOAD carriers of the MTHFR 677TT genotype [5, 21]. In this regard, previous studies by us in a subgroup of the samples included in the present study revealed increased plasma Hcy and decreased serum folate levels in LOAD patients than in controls, and the MTHFR 677TT genotype was associated with a significant reduction of serum folate levels that fostered an increase in plasma Hcy in LOAD individuals[11, 28].

Hyperhomocysteinemia (HHcy) promotes changes in DNA and protein methylation and oxidative stress, and contributes to Aβ₄₂ generation, inflammation, neurofibrillary tangle formation, and white matter damage and brain atrophy [2 , 30]. Indeed, several HHcy-related detrimental mechanisms could be implicated in neurodegeneration, including protein structural and functional modifications, oxidative stress, cellular metabolic derangements, epigenetic modifications, pathological aggregates deposition, endothelial damage, and atherothrombosis [31], so that it is often hard to excorporate Hcy levels from vascular risk and consequent worsening of predisposition to develop AD and other neurodegenerative pathologies. Indeed, HHcy has become the focus of interest in the research of several neurodegenerative disorders, including AD, Parkinson’s disease, and others, where it may be a precipitator/worsener of the susceptibility of the brain to the specific neurodegeneration [31], and the MTHFR C677T polymorphism has been associated to the risk of sporadic Parkinson’s disease, mild cognitive impairment, and vascular cognitive impairment, as well as to increased Hcy levels in those patients [32 –34]. Concerning the genetic model, it has been recently suggested that it is an additive model, with each carried MTHFR 677T allele increasing the amount of circulating Hcy [5].

In summary, the present results confirm data from recent meta-analyses suggesting that the MTHFR C677T polymorphism is also a LOAD risk factor in non-Asian populations and in APOE ɛ4 carriers. Furthermore, they are indicative that a contribution to LOAD risk in non-APOE ɛ4 carriers should not be excluded. The strengths of the present study are the relatively high sample size with respect to previous case-control studies in the field [13, 14], and the inclusion of only LOAD patients and controls of the same age. The limit of the study is mainly that, as explained in the Materials and Methods section, we could not genotype all the subjects for APOE. The search for genetic risk factors for LOAD is a timely issue, especially in non-APOE ɛ4 carriers, and MTHFR variants have been recently regarded among potential ones [35]. Confirmation of present data in larger non-APOE ɛ4 cohorts is required.

Footnotes

ACKNOWLEDGMENTS

Financial support was provided by the Italian Ministry of Health. The authors acknowledge Dr. Rosario Sebastiano Spada for the expert medical evaluation of patients at IRCCS Oasi Maria Santissima, Troina.

Authors’ disclosures available online ().

References

Shen

, Ji

(2015) Associations between homocysteine, folic acid, vitamin B12 and Alzheimer’s disease: Insights from meta-analyses. J Alzheimers Dis 46, 777–790.

Coppedè

(2010) One-carbon metabolism and Alzheimer’s disease: Focus on epigenetics. Curr Genomics 11, 246–260.

Seshadri

, Beiser

, Selhub

, Jacques

, Rosenberg

, D’Agostino

, Wilson

, Wolf

(2002) Plasma homocysteine as a risk factor for dementia and Alzheimer’s disease. N Engl J Med 346, 476–483.

Araújo

, Martel

, Borges

, Araújo

, Keating

(2015) Folates and aging: Role in mild cognitive impairment, dementia and depression. Ageing Res Rev 22, 9–19.

, Teng

, Li

, Hao

, Wang

(2016) Homocysteine and Alzheimer’s disease: Evidence for a causal link from Mendelian randomization. J Alzheimers Dis 52, 747–756.

Yamada

, Zhoutao

, Rima

, Mathews

(2001) Effects of common polymorphisms on the properties of recombinant human methylenetetrahydrofolate reductase. Proc Natl Acad Sci U S A 98, 14853–14858.

Rozen

(1997) Genetic predisposition to hyperhomocysteinemia: Deficiency of methelenetetrahydrofolate reductase (MTHFR). Thromb Haemost 78, 523–526.

Martinez-Frias

(2008) The biochemical structure and function of methylenetetrahydrofolate reductase provide the rationale to interpret the epidemiological results on the risk for infants with Down syndrome. Am J Med Genet A 146A, 1477–1482.

Nishio

, Goto

, Kondo

, Ito

, Ishida

, Kawai

, Naito

, Wakai

, Hamajima

(2008) Serum folate and methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism adjusted for folate intake. J Epidemiol 18, 125–131.

10.

Devlin

, Clarke

, Birks

, Evans

, Halsted

(2006) Interactions among polymorphisms in folate-metabolizing genes and serum total homocysteine concentrations in a healthy elderly population. Am J Clin Nutr 83, 708–713.

11.

Coppedè

, Tannorella

, Pezzini

, Migheli

, Ricci

, Ienco

, Piaceri

, Polini

, Nacmias

, Monzani

, Sorbi

, Siciliano

, Migliore

(2012) Folate, homocysteine, vitamin B12, and polymorphisms of genes participating in one-carbon metabolism in late-onset Alzheimer’s disease patients and healthy controls. Antioxid Redox Signal 17, 195–204.

12.

Zhang

, Miao

, Li

, Hu

(2010) Meta-analysis of the methylenetetrahydrofolate reductase C677T polymorphism and susceptibility to Alzheimer’s disease. Neurosci Res 68, 142–150.

13.

Peng

, Lao

, Huang

, Qin

, Li

, Zeng

(2015) The MTHFR C677T polymorphism contributes to increased risk of Alzheimer’s disease: Evidence based on 40 case-control studies. Neurosci Lett 586, 36–42.

14.

Rai

(2016) Methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism and Alzheimer disease risk: A meta-analysis. Mol Neurobiol, in press. doi: 10.1007/s12035-016-9722-8.

15.

, Zhao

, Zhang

, Zhan

(2009) Association of RFC1 A80G and MTHFR C677T polymorphisms with Alzheimer’s disease. Neurobiol Aging 30, 1601–1607.

16.

American

Psychiatric Association

(1994) Diagnostic and statistical manual of mental disorders, fourth ed., American Psychiatric Association. Washington, DC.

17.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’sdisease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

18.

Tannorella

, Stoccoro

, Tognoni

, Petrozzi

, Salluzzo

, Ragalmuto

, Siciliano

, Haslberger

, Bosco

, Bonuccelli

, Migliore

, Coppedè

(2015) Methylation analysis of multiple genes in blood DNA of Alzheimer’s disease and healthy individuals. Neurosci Lett 600, 143–147.

19.

Brunelli

, Bagnoli

, Giusti

, Nacmias

, Pepe

, Sorbi

, Abbate

(2001) The C677T methylenetetrahydrofolate reductase mutation is not associated with Alzheimer’sdisease. Neurosci Lett 315, 103–105.

20.

Religa

, Styczynska

, Peplonska

, Gabryelewicz

, Pfeffer

, Chodakowska

, Luczywek

, Wasiak

, Stepien

, Golebiowski

, Winblad

, Barcikowska

(2003) Homocysteine, apolipoproteine E and methylenetetrahydrofolate reductase in Alzheimer’s disease and mild cognitive impairment. Dement Geriatr Cogn Disord 16, 64–70.

21.

Roussotte

, Narr

, Small

, Thompson

, Alzheimer’s Disease Neuroimaging Initiative (2016) The C677T variant in MTHFR modulates associations between blood-based and cerebrospinal fluid biomarkers of neurodegeneration. Neuroreport 27, 948–951.

22.

Bertram

, McQueen

, Mullin

, Blacker

, Tanzi

(2007) Systematic meta-analyses of Alzheimer disease genetic association studies: The AlzGene database. Nat Genet 39, 17–23.

23.

Laumet

, Chouraki

, Grenier-Boley

, Legry

, Heath

, Zelenika

, Fievet

, Hannequin

, Delepine

, Pasquier

, Hanon

, Brice

, Epelbaum

, Berr

, Dartigues

, Tzourio

, Campion

, Lathrop

, Bertram

, Amouyel

, Lambert

(2010) Systematic analysis of candidate genes for Alzheimer’s disease in a French, genome-wide association study. J Alzheimers Dis 20, 1181–1188.

24.

Naj

, Beecham

, Martin

, Gallins

, Powell

, Konidari

, Whitehead

, Cai

, Haroutunian

, Scott

, Vance

, Slifer

, Gwirtsman

, Gilbert

, Haines

, Buxbaum

, Pericak-Vance

(2010) Dementia revealed: Novel chromosome 6 locus for late-onset Alzheimer disease provides genetic evidence for folate-pathway abnormalities. PLoS Genet 6, e1001130.

25.

Ren

, Wang

, Fang

, Liu

, Wang

, Tang

, Deng

, Xu

, Wang

, Chen

(2011) The MTHFD1L gene rs11754661 marker is associated with susceptibility to Alzheimer’s disease in the Chinese Han population. J Neurol Sci 308, 32–34.

26.

, Yu

, Wu

, Zhang

, Liu

, Wang

, Tan

(2012) Replication of the MTHFD1L gene association with late-onset Alzheimer’s disease in a Northern Han Chinese population. J Alzheimers Dis 29, 521–525.

27.

Wakutani

, Kowa

, Kusumi

, Nakaso

, Yasui

, Isoe-Wada

, Yano

, Urakami

, Takeshima

, Nakashima

(2004) A haplotype of the methylenetetrahydrofolate reductase gene is protective against late-onset Alzheimer’s disease. Neurobiol Aging 25, 291–294.

28.

Coppedè

, Grossi

, Buscema

, Migliore

(2013) Application of artificial neural networks to investigate one-carbon metabolism in Alzheimer’s disease and healthy matched individuals. PLoS One 8, e74012.

29.

Smith

, Refsum

(2016) Homocysteine, B vitamins, and cognitive impairment. Annu Rev Nutr 36, 211–239.

30.

Grossi

, Stoccoro

, Tannorella

, Migliore

, Coppedè

(2016) Artificial neural networks link one-carbon metabolism to gene-promoter methylation in Alzheimer’s disease. J Alzheimers Dis 53, 1517–1522.

31.

Bonetti

, Brombo

, Zuliani

(2016) The relationship between hyperhomocysteinemia and neurodegeneration. Neurodegener Dis Manag 6, 133–145.

32.

Rajagopalan

, Jahanshad

, Stein

, Hua

, Madsen

, Kohannim

, Hibar

, Toga

, Jack

Jr. , Saykin

, Green

, Weiner

, Bis

, Kuller

, Riverol

, Becker

, Lopez

, Thompson

, Alzheimer’s Disease Neuroimaging Initiative (ADNI), Cardiovascular Health Study (CHS) (2012) Common folate gene variant, MTHFR C677T, is associated with brain structure in two independent cohorts of people with mild cognitive impairment. Neuroimage Clin 1, 179–187.

33.

Skrobot

, McKnight

, Passmore

, Seripa

, Mecocci

, Panza

, Kalaria

, Wilcock

, Munafò

, Erkinjuntti

, Karhunen

, Pessi

, Martiskainen

, Love

, Genetic and Environmental Risk for Alzheimer’s disease Consortium (GERAD1), Kehoe

(2016) A validation study of vascular cognitive impairment genetics meta-analysis findings in an independent collaborative cohort. J Alzheimers Dis 53, 981–989.

34.

Zhu

, Zhu

, He

, Liu

(2015) Association of MTHFR C677T with total homocysteine plasma levels and susceptibility to Parkinson’s disease: A meta-analysis. Neurol Sci 36, 945–951.

35.

Román

(2015) MTHFR gene mutations: A potential marker of late-onset Alzheimer’s disease? J Alzheimers Dis 47, 323–327.